MAY1967
FLUORIDE IONCATALYZED HYDROLYSIS OF CARBOXYLIC ANHYDRIDES
1547
Fluoride Ion Catalyzed Hydrolysis of Carboxylic Anhydrides’ C. A. BUNTONAND J. H. FENDLER Department of Chemistry, University of California at Santa Barbara, Santa Barbara, California Received November 10, 1966 Fluoride ion will catalyze the hydrolysis of acetic and succinic anhydride in water and in aqueous dioxane. The ratios of the rates of the catalyzed to uncatalyzed reactions are larger in aqueous dioxane than in water. For the aqueous hydrolysis of acetic anhydride the rate of the catalyzed disappearance of anhydride is greater than the rate of appearance of acetic acid, showing that acetyl fluoride can build up in concentration, nnd can be aspirated out of the solution. Fluoride ion acts predominantly as a nucleophilic catalyst. Sodium hydrogen fluoride and hydrofluoric acid also catalyze the hydrolysis. The hydrolysis of succinic anhydride is catalyzed by fluoride ion, but in this system acyl fluoride does not build up in concentration.
Hydrolyses of carboxylic anhydrides and esters are catalyzed both by nucleophiles and by general basesa2 A nucleophile can convert the substrate into a more reactive acyl compound; examples in anhydride hydrolysis are the formate and nitrite ion catalyzed hydrolyses of acetic anhydride, where acetyl formate and nitrite are intermediates. 3,4 General base catalysis could remove a proton from a n attacking water molecule, or from a tetrahedral intermediate formed by addition of a water molecule to the ~ u b s t r a t e . ~ , ~ It is often difficult to differentiate between these different reaction paths because basicity and nucleophilicity toward a carbonyl center are related. Tests based on the magnitude of the solvent deuterium isotope effect have been used, but are not completely reliable.’,* Bell has noted that small basic anions of high charge density are often very effective as general base catal y s t ~and , ~ fluoride ion must act as a base rather than a nucleophile in catalyzing the hydrolysis of acetyl fluoride. l o Fluoride ion can also act as a nucleophile. It is not particularly reactive toward alkyl groups,” but it is highly reactive toward acyl centerP and to aryl sulf i t e ~ . (In ~ ~ some of these reactions nucleophilic and general base catalysis were not separated.)
Experimental Section Kinetics.-Reaction was followed either titrimetrically, by following the formation of acetic acid, or by following the disappearance of acetic anhydride spectrophotometrically. These general methods have been described el~ewhere.1~ (1) (a) This work was supported by the National Science Foundation under Grant GP 2889. This support is gratefully acknowledged. (h) Part of this work was presented at the Symposium on Physical Organic Chemistry, 151st National Meeting of the American Chemical Society, Atlantic City, N . J., Sept 1965. (2) For a general review, see T . C. Bruice and 8. Benkovic, “BioOrganic hfechanisms,” W. T. Benjamin Ino., New York, N. Y., 1966, Chapter I. (3) M . Kilpatrick, J. Am. Chem. SOC.,SO, 2891 (1928); V. Gold and E. G. Jefferson, J. Chem. Soc.. 1416 (1953). (4) E. B . Lees and B. Saville, ibid., 2262 (1958). (5) A. R . Butler and V. Gold, ihid., 2305 (1961); 2212 (1962). (6) C . -4. Bunton and V. J. Shiner, J . Am. Chem. Soc., 83, 3207 (1961); S.L. Johnson, ihid., 84, 1729 (1962); C. A. Bunton, N. A. Fuller, 5. G. Perry, and V. J . Shiner, J . Chem. Soc., 2918 (1963). (7) M . L. Bender, E. J. Pollock, and M . C. BlcNeveu, J . Am. Chem. Soc., 8 4 , 5 9 5 (1962). (8)D. G . Oakenfull, T. Riley, and V. Gold, Chem. Commun., 385 (1966). (9) R. P. Ilell, “The Proton in Chemistry,” Cornell University Press, Ithaea, K.Y., 1959, p 179. (IO) C. A . IIunton and J. H. Fendler, J . Org. Chem., 31, 2307 (1966). (11) R . F. Hudson, Chimia (Aarau), 16, 173 (1962). (12) M. Green and R. F. Hudson, J . Chem. Soc., 1055 (1962); W. P. Jencks and J. Carriuolo, J. Am. Chem. Soc., 83, 1743 (1961); J. 0. Edwards and R . G . Pearson, ibid., 84, 16 (962). (13) C . A . Bunton and G . Schwerin, J. Org. Chem., 31, 842 (1966).
The appearance of acetic acid can be followed most conveniently by use of a pH-Stat, with disodium hydrogen phosphate as titrant. Under these conditions with this titrant there is no enhanced hydrolysis of any acetyl fluoride which may be formed as an intermediate,Io so that for a possible two-stage reaction (CHaC0)zO
1 2 + F- + CHaCOz + CH3COF + Hz0
CHaCOzH
+ HF
it is the conversion of acylating agents (acetic anhydride plus acetyl fluoride) to acid which is being followed. Experiments on acetyl fluoride itself showed that it was not lost by evaporation during hydrolysis followed in a pH-Stat. We made preliminary measurements on the rate of formation of acetic acid by adding portions of the reaction solution to an excess of aniline, which reacts rapidly with the acetylating agents as shown below, and then titrating the residual acid with dilute sodium hydroxide under nitrogen, using phenolphthalein as indicator.’* This method is not particularly satisfactory for (CH3CO)zO
HZO + F- +CHICOF + CHaCOz- +
CHaCOzH
+ HF
fast
+ CsHsNHz +CHICOZH + CHaCONHCsHs fast CH3COF + CsHsNHz +H F + CHaCONHCaHS
(CHaC0)zO
following the fluoride ion catalyzed hydrolysis of acetic anhydride. There is a risk that the volatile acetyl fluoride would be lost during sampling, and the reaction rate is such that only the latter part of the reaction could be followed. However, this “aniline” method appears to be satisfactory for the slower fluoride ion catalyzed hydrolysis of succinic anhydride, and for hydrolyses in aqueods dioxane. For hydrolysis followed by the “aniline” method 0.05 ilf anhydride was used, and the pH-Stat runs were carried out a t pH 5.5 with 0.05 M acetic anhydride. The extinction coefficient of acetic anhydride, measured in ethanol rather than water, because of the reactivity of the anhydride, is 61 at 2450 A . At this wavelength the extinction coefficient of acetic acid is very low, and when acetyl fluoride, sufficient to give an 0.01 ilf solution, was injected into water, we observed no increase in absorbance a t 0” within 5 sec, during which time there would have been little hydrolysis.10 The conversion of acetic anhydride into either acetyl fluoride or acetic acid is followed spectrophotometrically, and by this method we should follow stage one of a hypothetical two-stage reaction. We used either a Cary 14 or a Gilford 220 spectrophotometer with thermostated cell compartments and added the anhydride to a cell which contained the reagents a t the reaction temperature.“ All hydrolyses in aqueous dioxane were followed titrimetrically by addition of portions of the reaction mixture to aniline, and subsequent acid-base titration. The hydrolysis of succinic anhydride was followed either spectrophotometrically, or by the “aniline” method. The integrated first-order rate constants were determined graphically. Aspiration Experiments.-An amount of acetic anhydride, sufficient to give a 0.1 M solution, was dropped into an aqueous (14) C. A. Bunton and S. G . Perry, J . Chem. Soc., 3070 (1960); C. A. Bunton and J. H. Fendler, J . Org. Chem., 30, 1365 (1965).
BUNTON AND FENDLER
VOL.32
I5
P I
a
P J;
IO
0
3
2
4
6
8
51
Time, min.
Figure 1.-Plots of log (titer, - titert) (solid points) and log [Ac,O], (open points) against time, for reaction of acetic anhydride in 0.3 M sodium fluoride a t OD. The straight line drawn approximately through the solid points corresponds to kt = 39.0 X l o d 4sec-l.
I
0.5
1.0
I.5
IF], M . solution of sodium or potassium fluoride a t 0” which had been saturated with nitrogen gas, and which was in a bubbler with a sintered-glass frit, so that the bubbles swept through all the solution. The bubbler was shaken, and a stream of nitrogen was passed through the solution. After complete reaction the acetic acid in the solution was titrated with dilute alkali, and from this titration value and the amount of acetic anhydride initially added we could calculate how much volatile material had been swept out. Control tests with a solution of acetic acid showed that
